Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/148325
Full metadata record
DC FieldValueLanguage
dc.contributor.authorQiu, Hanen_US
dc.contributor.authorZheng, Qinkaien_US
dc.contributor.authorZhang, Tianweien_US
dc.contributor.authorQiu, Meikangen_US
dc.contributor.authorMemmi, Gerarden_US
dc.contributor.authorLu, Jialiangen_US
dc.date.accessioned2021-05-05T08:43:52Z-
dc.date.available2021-05-05T08:43:52Z-
dc.date.issued2021-
dc.identifier.citationQiu, H., Zheng, Q., Zhang, T., Qiu, M., Memmi, G. & Lu, J. (2021). Toward secure and efficient deep learning inference in dependable IoT systems. IEEE Internet of Things Journal, 8(5), 3180-3188. https://dx.doi.org/10.1109/JIOT.2020.3004498en_US
dc.identifier.issn2327-4662en_US
dc.identifier.other0000-0003-2678-8070-
dc.identifier.other0000-0002-5391-9446-
dc.identifier.other0000-0001-6595-6650-
dc.identifier.other0000-0002-1004-0140-
dc.identifier.other0000-0002-3380-8394-
dc.identifier.other0000-0002-6752-7224-
dc.identifier.urihttps://hdl.handle.net/10356/148325-
dc.description.abstractThe rapid development of deep learning (DL) enables resource-constrained systems and devices [e.g., Internet of Things (IoT)] to perform sophisticated artificial intelligence (AI) applications. However, AI models, such as deep neural networks (DNNs), are known to be vulnerable to adversarial examples (AEs). Past works on defending against AEs require heavy computations in the model training or inference processes, making them impractical to be applied in IoT systems. In this article, we propose a novel method, Super-IoT, to enhance the security and efficiency of AI applications in distributed IoT systems. Specifically, Super-IoT utilizes a pixel drop operation to eliminate adversarial perturbations from the input and reduce network transmission throughput. Then, it adopts a sparse signal recovery method to reconstruct the dropped pixels and wavelet-based denoising method to reduce the artificial noise. Super-IoT is a lightweight method with negligible computation cost to IoT devices and little impact on the DNN model performance. Extensive evaluations show that it can outperform three existing AE defensive solutions against most of the AE attacks with better transmission efficiency.en_US
dc.language.isoenen_US
dc.relationCHFA-GC1-AW03en_US
dc.relation.ispartofIEEE Internet of Things Journalen_US
dc.rights© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/JIOT.2020.3004498en_US
dc.subjectEngineering::Computer science and engineeringen_US
dc.titleToward secure and efficient deep learning inference in dependable IoT systemsen_US
dc.typeJournal Articleen
dc.contributor.schoolSchool of Computer Science and Engineeringen_US
dc.identifier.doi10.1109/JIOT.2020.3004498-
dc.identifier.scopus2-s2.0-85101681631-
dc.identifier.issue5en_US
dc.identifier.volume8en_US
dc.identifier.spage3180en_US
dc.identifier.epage3188en_US
dc.subject.keywordsInternet of Thingsen_US
dc.subject.keywordsSensorsen_US
item.fulltextWith Fulltext-
item.grantfulltextopen-
Appears in Collections:SCSE Journal Articles
Files in This Item:
File Description SizeFormat 
paper.pdf1.49 MBAdobe PDFView/Open

SCOPUSTM   
Citations 20

5
Updated on May 17, 2022

PublonsTM
Citations 20

6
Updated on May 11, 2022

Page view(s)

181
Updated on Jul 6, 2022

Download(s) 50

34
Updated on May 17, 2022

Google ScholarTM

Check

Altmetric


Plumx

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.